Test AI Agents in the Playground

TL;DR

The Testing Playground lets you preview an AI Agent's reasoning and responses before deployment - free of usage costs and without affecting live items. Test with sample items, refine prompts, and iterate freely until satisfied, then click Create to activate.

Overview
Why Use the Playground
How to Test in the Playground
Best Practices
What’s Next?
 

Overview

The Testing Playground lets you preview exactly how an AI agent will reason and respond before deploying it. This helps you validate behaviour and refine prompts without affecting real work items.

Why Use the Playground

• See the AI agent's reasoning - understand why the agent makes each decision.
• No usage cost - Playground runs do not count toward your AI usage limits, so you can iterate freely on prompts and configurations before deploying the agent.
• Catch issues early - confirm the AI agent behaves as expected before it acts on live items.

How to Test in the Playground

1. During agent setup (after configuring triggers and actions), open the Testing Playground.
2. Select sample items to run the AI agent against.
3. Review how the agent reasons and responds, including its full reasoning process.
4. Refine your prompts and configuration as needed.
5. Re-run as many times as necessary - iterate freely.
6. When satisfied, click Create to activate your AI agent.

Best Practices

• Test first: Always validate your AI agent's behaviour in the Playground before deploying.
• Use real examples: When troubleshooting, test with the actual item that produced an unexpected result.
• Set fallback values for cases where the AI agent can't make a determination.
• Start small: Test one AI agent type in a single folder before expanding.

What’s Next?

• Write Effective Agent Prompts
• AI Agent Limitations and Troubleshooting